International J.of Multidispl.Research & Advcs. in Engg.(IJMRAE),
ISSN 0975-7074, Vol. 2, No. I, April 2010, pp 167-180
PERSON IDENTIFICATION USING IRIS RECOGNITION BASED
ON DCT AND NEURAL NETWORK
Design and evaluation of Iris recognition system for person identification is discussed in this paper.
As technology advances information and intellectual properties are wanted by many unauthorized
personnel. As a result many organizations are searching ways for more secure authentication methods
for the user access. In network security there is a vital emphasis on the automatic personal
identification. Due to its inherent advantages biometric based verification especially iris identification
is gaining a lot of attention. Iris recognition uses iris patterns for personnel identification. The system
steps are capturing iris image, localizing iris and the iris pattern recognition. The iris is extracted from
the eye image. Due to the high degree of freedom in iris pattern only part of the iris structure is
selected for recognition. The proposed method is Discrete Cosine Transform (DCT) coefficient based
technique that extracts important features using transformed coefficients. Obtained features are fed to
generalized feed forward neural network with different learning rules and activation functions for
person identification. Experimental results show that Discrete Cosine Transform (DCT) based feature
extraction technique has an encouraging performance.
Keywords: Biometrics, Iris recognition, Personal identification, DCT, FFT, ANN, MLP.